Closed-loop critical results communication in teleradiology

The most expensive study your facility will ever handle is one that was read correctly and told to no one. The hemorrhage is in the impression, the report is signed inside its turnaround window, and the patient spends the next several hours deteriorating while the answer sits unread in the chart. Closed-loop critical results communication exists to close that gap: the critical finding is verbally conveyed to a licensed provider who can act on it, and the loop closes only when that provider's name, the read-back, and the timestamp are recorded. The Joint Commission requires each organization to define which results are critical, who reports them, to whom, and within what timeframe, and then to measure its own compliance against that definition.

Teleradiology adds distance to every step of that handoff. Most vendor service descriptions compress the whole subject into one sentence about calling the referring physician, which is why buyers shopping for teleradiology with critical results communication spend an afternoon comparing near-identical sentences. This page covers the standards that govern the obligation, what the process contains, where remote reading breaks it, the escalation ladder that runs when nobody answers at 3 a.m., the contract language that makes a vendor's obligation enforceable, and the log format a quality committee will ask you for.

Which standard requires closed-loop reporting, exactly?

Two documents govern critical results reporting in a US imaging program, and neither one hands you a number. The Joint Commission's National Patient Safety Goal on critical results, carried for years as NPSG.02.03.01, requires the organization to report critical results of tests and diagnostic procedures on a timely basis and makes every definition behind that sentence the organization's own work.

Read the goal element by element and it asks for four things: a written definition of which results and values are critical, a written statement of by whom and to whom they are reported, an acceptable length of time between the availability of a result and its receipt by the responsible licensed caregiver, and measurement of the organization's own compliance with that definition. Hospitals and critical access hospitals saw the goal renumbered NPG.01.02.01 effective January 1, 2026, with the substance carried across intact. Cite whichever number your accreditation manual uses, and write your vendor contract against the four elements so a renumbering never reopens the agreement.

The ACR practice parameter for communication of diagnostic imaging findings, published on acr.org and revised in 2020, covers the interpreting physician's side of the same duty. It describes non-routine communication as communication that is timely enough to permit prompt action, delivered in a manner that reasonably ensures receipt, and documented in the final signed report. It is a practice parameter and carries no survey authority, which understates its weight: it is the document that gets read aloud when a communication failure reaches litigation.

Neither text specifies 30 minutes. If your policy carries that number, your policy invented it, and a surveyor will hold you to it exactly as written. The Joint Commission accredits your facility, so the demonstration is yours to make. AstraRad holds no Joint Commission accreditation, and no vendor accreditation would substitute for your own.

Your facility defines the critical results list

The critical results list belongs to your facility, and your vendor reads against it. That is the practical consequence of the four elements above: the goal fixes the obligation and leaves you the content. The ACR parameter tiers the same territory, separating findings that suggest a need for immediate or urgent intervention from findings that are significant and unexpected and from findings discrepant with a preceding interpretation of the same examination.

Most facilities land on three categories. Emergent findings, the acute intracranial hemorrhage and aortic dissection tier, get immediate voice contact, typically within 30 to 60 minutes of signature. Urgent findings, a new suspicious mass or an outpatient DVT, get direct contact with an acknowledgement inside the same shift or the same day. Discrepant or unexpected findings, including material disagreement with a prior interpretation, get a defined window, often 24 to 72 hours.

The gap between your list and a vendor's default list stays invisible until it hurts. A national operation arrives with a critical findings list built from its own book of business. Maybe your policy carries a category their radiologists were never told to call on, or your emergent window is 30 minutes and theirs is an hour. Nobody catches it at go-live because both lists look reasonable, and the studies that would expose the difference are rare. Send your list before the first study moves, in writing, and require written confirmation that the panel reads against your list as amended.

Five data points close a critical results notification

A critical results notification is closed when five things are true and recorded. Anything less is a delivery.

Data point	What it proves
Study and finding, with the assigned category	What triggered the obligation, and under which tier
Name and role of the licensed provider who received it	A human authorized to act took the handoff
Name of the radiologist who conveyed it	The communication ran physician to physician
Timestamps for report signature and successful contact	The elapsed interval you must measure against your own policy
Confirmation of receipt, and the method used	The read-back. This is the element that closes the loop

Mirror those five fields in the final signed report itself, as a communication statement naming the receiving clinician's role and the time of contact. Two records, one in the report and one in the log, means the evidence survives a system migration or a vendor change.

The test to run in any vendor demo is simple. Ask for a month of critical results entries as an export and check every row for all five fields. A row that reads "called to ED" with a date and no receiving name won't hold up under survey.

Where does critical results communication break down when the radiologist is remote?

Remote reading breaks the loop in three places: the call that never happens, the call that reaches an answering service, and the log that stays in the vendor's system.

The first failure is the silent one, and it deserves the most attention because it leaves no trace. The impression contains unambiguous critical language. The study came back inside its turnaround target. Everyone behaved reasonably: the radiologist assumed the flag in the reporting system would surface on the facility side, the facility assumed the vendor's workflow included a call, and the ED had already moved to the next patient. When the chart is reviewed later it looks complete, because a chart records what happened, and this failure consists entirely of something that didn't. There's no misdialed number to find, no voicemail to replay, no exception report to pull. The only way to catch it prospectively is to reconcile the vendor's critical results log against every report that carried critical language, month by month, and ask why an emergent finding has no matching notification row. Facilities that run that reconciliation tend to find gaps within the first quarter. Facilities that skip it find them during a survey or a deposition.

The second failure is the answering-service relay. Contact runs through a toll-free number, the message is left with whoever picks up, and the log entry reads "message left for ED." A relay through a non-licensed intermediary produces no read-back from a provider who can act on the finding. Clinicians name this as their most frequent complaint about large remote reading operations, and it's the point to probe hardest in a reference call.

The third failure is a log split across two organizations. The Joint Commission asks you to demonstrate compliance against your own defined timeframe under NPG.01.02.01, the critical-results goal that replaced NPSG.02.03.01 for hospitals and critical access hospitals on January 1, 2026. When the log lives in the vendor's system, demonstrating it depends on that vendor producing a report on request. If the relationship ends, that history is hard to retrieve.

The same distance also runs backward. A referring physician who wants the reading radiologist on the phone at 3 a.m. for a clarifying question needs a direct route, and a panel you can't reach for a consult won't reliably reach you either. Direct radiologist access covers what that route has to look like in a contract.

Four contact attempts, then a documented exception

A hemorrhage signed at 3:04 a.m. and an ordering physician whose direct line rings out is the scenario most vendor service descriptions skip, and it is the one your policy has to answer before it happens. When the first call goes unanswered, a workable emergent ladder runs four attempts and ends in a documented exception.

Each step names a role your facility staffs overnight, a wait before escalation, and a log entry that survives the shift.

Attempt	Who the radiologist reaches for	Wait before escalating	What the log records
1	Ordering provider, direct line	3 minutes	Time dialed, number used, outcome
2	Covering provider, or the unit's charge nurse who routes to a licensed provider	3 minutes	Role reached, time, whether a licensed provider took the finding
3	ED attending on duty, or the house supervisor	5 minutes	Role reached, time, acknowledgement or failure
4	Administrator on call	Ladder ends	Failed-contact exception, time raised, named facility recipient

Three rules make the ladder function. The roster is yours and it has to be current: contact lists go stale within a quarter, so set a refresh cadence and hold the vendor to the latest version. A failed contact generates a documented exception, never silence. And the escalation clock starts at report signature, because "when someone remembered to start dialing" can't be measured.

Urgent, non-emergent categories can run a longer ladder, and non-voice channels with a defined acknowledgement window are usually acceptable there. Write the per-category distinction down. A radiologist covering your account overnight shouldn't be improvising tier policy.

What does the clock do while you escalate at 3 a.m.?

Nothing stops it. The notification interval runs from report signature to the moment a licensed provider acknowledges the finding, so every unanswered ring, every hold, and every transfer to a unit phone is elapsed time inside your own window. A ladder that takes 14 minutes to reach an ED attending has spent 14 minutes of a 30-minute policy window, and compliance is scored on the outcome.

Four clock rules keep the measurement honest, and each one closes a way the interval quietly gets shorter than it was.

• The clock starts at signature, never at the first dial. "When someone got around to calling" is unmeasurable, and it is the field vendors most often supply in place of the signature timestamp.
• Every attempt carries its own timestamp. A single contact time on a row that took four attempts hides the ladder, and hiding the ladder is what makes an exception unauditable.
• Handoffs do not reset it. If the signing radiologist reaches the end of a shift mid-escalation, the obligation transfers with the finding and the interval keeps accumulating against the original signature.
• An exception stops the clock only by being raised. The row closes when a named person at your facility receives the failed-contact exception, with that delivery timestamped. A message left in a queue closes nothing.

One substitution rule belongs alongside them. When acknowledgement comes from someone other than the ordering clinician, which is the normal outcome overnight, the log records who took it, their role, and why the substitution happened. That single field is what lets you answer the only question a reviewer asks about a 3 a.m. call: did a person authorized to act on this finding actually hear it?

Put the critical results obligations in the contract

Critical results obligations belong in the agreement itself, in plain language you can hold someone to. Start with an exhibit that states the notification window for each of your categories, because every clause below refers back to it.

Category	Typical findings	Notification window to write into the exhibit	Channel the clause should require
1, emergent	Acute intracranial hemorrhage, aortic dissection, tension pneumothorax	Acknowledgement inside your policy window, commonly 30 to 60 minutes from signature	Direct voice contact with a licensed provider
2, urgent	New suspicious mass, outpatient DVT, unexpected findings needing action within the day	Acknowledgement inside the same shift or the same day	Voice preferred, secure message acceptable with an acknowledgement deadline
3, discrepant or unexpected	Material disagreement with a prior interpretation, incidental findings requiring follow-up	Commonly 24 to 72 hours	Secure message or voice, acknowledgement recorded either way

The windows in that middle column are the ones most facilities land on. Yours govern, and the clause should say so in the words "as stated in Exhibit A and amended by Client from time to time." Ten provisions then carry the weight.

Who calls. "The interpreting radiologist shall personally place the notification for Category 1 findings. Notification by a call center, answering service, or non-physician staff member does not satisfy this obligation."

What counts as contact. "Communication shall be by direct voice contact with a licensed provider authorized to act on the finding. Voicemail, secure message, fax, portal notification, or a message left with non-licensed staff leaves the notification open."

Which list governs. "Vendor shall interpret against Client's critical results definitions and category timeframes as amended from time to time, and shall confirm receipt of each amendment in writing."

Attempts and escalation. "Vendor shall follow Client's escalation roster with no fewer than three documented attempts for Category 1 findings, and shall record a failed-contact exception delivered to Client's designated contact within 15 minutes of the final attempt."

The log. "Vendor shall deliver a monthly critical results log in machine readable format containing, for each notification: study identifier, finding category, receiving provider name and role, interpreting radiologist name, report signature timestamp, contact timestamp, and confirmation of receipt."

Measured compliance. "Vendor shall report monthly the percentage of Category 1 notifications closed within Client's defined timeframe." Most contracts omit this clause, and it's the one that maps directly onto what you must demonstrate about your own program.

Reachability in both directions. "Vendor shall make the interpreting radiologist available for consultation by direct call-back within 15 minutes during the shift in which the report was signed."

Windows by tier. "Vendor shall obtain documented acknowledgement within the notification window stated in Exhibit A for each finding category, measured from report signature to confirmed receipt by a licensed provider, and shall apply the amended Exhibit within five business days of Client's written notice."

The call follows the interpretation. "Where Vendor subcontracts or reassigns an interpretation to any affiliate, overflow panel, or third party, the notification obligation follows the interpretation and Vendor remains responsible to Client for its performance. Each subcontracted read shall be identified as such in the critical results log."

Audit evidence. "On 10 business days' notice, Vendor shall produce, for any period within the preceding 24 months, the complete critical results log, the per-attempt contact records behind any notification Client identifies, every failed-contact exception raised with its delivery timestamp and recipient, and the escalation roster version in force on the date of each notification. This obligation survives termination for the duration of Client's record retention policy."

That last clause is the one to fight for, because four artifacts answer nearly every question a surveyor or an attorney will put to you: the log, the per-attempt records, the exceptions, and the roster version. A vendor who can hand over the log but none of the other three can show you a result and never the process that produced it, which is exactly the evidence gap that closed-loop reporting exists to prevent.

If a vendor balks at the "who calls" clause and the reachability clause, you're buying report delivery, not closed-loop communication. One honest caveat: if your studies are read in-house 24/7 and your radiologists share a hallway with your ED, most of this contract language solves a problem you don't have. It matters when the reading physician and the treating physician will never meet.

What the critical results log looks like in a monthly quality report

A quality committee does not want the raw feed, and it will not read 400 rows. It wants one page: a count, a compliance percentage against your own window, the exceptions, and the rows standing behind both. Structure the monthly extract in three blocks and the committee packet assembles itself.

The summary block is the committee slide: notifications by category, the percentage closed inside your defined window, the median interval from report signature to acknowledgement, and the count of failed-contact exceptions. Four figures, each computed from the rows below it, each stated with its denominator. This is the block that answers the measurement element of the safety goal, so it is the block a surveyor asks to see first.

The row detail is one line per notification, in a format your quality staff can sort and filter.

Column	Contents	Why the committee asks for it
Study	Accession identifier and modality	Ties the row to the imaging record
Category	1, 2, or 3 under your own definitions	Determines which window applies
Signed at	Report signature timestamp	Starts the clock
Acknowledged at	Timestamp of confirmed receipt	Stops the clock
Interval	Minutes between the two	The measured number the percentage is built from
Received by	Name and role of the licensed provider	Proves a person authorized to act took it
Conveyed by	Interpreting radiologist	Proves the handoff ran physician to physician
Attempts	Count, and the ladder step reached	Shows the escalation actually ran
Outcome	Closed, or exception raised with its recipient	Separates a closed loop from an open one

The exception appendix gets a page per failed contact: what was tried, when, who at your facility was notified, and what happened to the patient's care afterward. Exceptions should be rare, which is precisely why a committee reads every one of them line by line.

Ask for the whole extract in a machine readable file alongside the PDF. A packet you cannot filter is a packet you cannot reconcile against your own report archive, and that reconciliation is the check that catches the notification that never happened. The same discipline applies to interpretation quality, which is why the critical results extract belongs next to the discrepancy data in a single review; how to audit radiology report quality covers the other half of that packet.

AstraRad puts the interpreting radiologist on the phone

The AstraRad subspecialist who signed the report places the critical results call and stays with it until a licensed provider acknowledges the finding. The panel is staffed in shifts 24/7/365, so a study that finishes at 3 a.m. at your facility lands with a radiologist who is awake and mid-shift. Every one of them is board-certified, fellowship-trained, physically located in the United States, and licensed in the state where your patients are located. Every report is a final signed report. Turnaround tiers are measured from last-image arrival to radiologist signature, and the full methodology is published on our SLA and quality page. The coverage model is described in how we work.

Four commitments sit behind the call. We read against your critical findings list and your escalation roster, refreshed on a stated cadence, with a documented failed-contact exception routed to your designated contact when the ladder runs out. Every notification is logged with the five fields above, delivered to you in exportable form, and mirrored as a communication statement in the final signed report. QA findings enter the same loop: our protocol gives 1 in 20 studies an independent blinded double-read, and a QA finding that meets your critical criteria triggers the same closed-loop call. And the interpreting subspecialist takes direct call-backs during the shift the report was signed; our radiologists explains how subspecialty assignment works.

Results move over HL7 and FHIR, images over DICOM from your PACS or portal upload, under the HIPAA aligned program documented on our compliance page.

If critical results handling is why you're evaluating overnight coverage, pair this page with STAT CT reads and STAT radiology reads, which cover the turnaround side of the same workflow. And because a preliminary interpretation that gets revised creates a Category 3 notification obligation of its own, read STAT vs preliminary vs final reads before you accept a preliminary-read model.

Pricing is per report, with no minimums or platform fees. If critical results communication is the clause you most need to get right, book a coverage consultation and send your critical findings policy and escalation roster. We'll map our workflow against both, clause by clause, before you sign anything. The report is the evidence. The call is the care.